I just came here to post that I couldn't read past 'The complexity was real, but distributed'. I can get past these LLM constructions when Claude uses them in chat. They seriously undermine the credibility of comment pieces or guides like this one when I encounter them. I absolutely hate it when I get them in a lengthy response to a simple question to a colleague about why we're going to do something a particular way.
I can't tell you how many times a week claude opus 4.8 high effort has to apologise for being wrong when I'm asking it about something narrow and specific that i want it to research but it blurts out broad context from its training material and incorrect conclusions/assumptions. This is happening all the time. Someone needs to create a repository of its apologies to remind us all of its limitations.
There's a huge difference between say, weilding a hatchet on a camping trip, and trying to get the hang of a splitting axe, with a 3ft or longer handle, when you're a kid. Getting a long, sweeping arc that comes down in the right spot isn't easy and if the axe head's centrifugal force pulls it away from you, you clunk the handle down on the wood. I definitely recall my hands ringing and numb from those kinds of impact. I don't remember ruining a handle, but if it had been my chore, I think I could've come close.
No, it's not a bunch of instructions, it's a colossal array of vectors that are the outcome of many thousands of lifetimes' worth of stimuli and reinforcement - not dissimilar in (very) abstract terms to the neurons in our brains.
I just meant that if I value, say, the security afforded by land that's been cleared of humans, and between them my owned AIs can deliver this as a 'surplus' outcome, just by processing resources available at low or no cost (sunlight, atmospheric gases, processed carcasses, etc.) then letting them run amuck is still 'economically valuable' to me, but doesn't require any human input, or a human counterparty. Very reductive example, but you can imagine a much more complex economy where a multitude of similar arrangements interact/compete to deliver outcomes of this type.
You couldn't have. Mechanical machines couldn't organise themselves into human-free supply chains that are economically productive for the owners of capital. AIs could.
And the time wasted on transfers. I used to regularly fly from airports in NY, London, SF, Singapore, LA, Sydney, etc. I would block out the opportunities to work or rest, and the reality was that only the plane time was valuable for either. It was painful to see all the other blocks of non productive time, particularly the allowances for congestion and disruption between downtown and the airports. I would have paid thousands a flight to be able to check in/clear security at my hotel and then get driven to a holding bay at the airport, and then on to the gate, in a vehicle suitable for both work and rest.
No one seems to care anymore, but a big issue that people were concerned about in the 2000s was the switch from 'I know more about me than the blob (corps, gov, etc) does' to, 'I need the blob to remind me where the hell I was that day'. Heart rate and blood oxygen data are hard to exploit data points but not impossible(1), but facing an accusation from someone who knows more about your movements than you do is an uncomfortable scenario. Of course right now, if you're facing an acusation of this type, odds are it's legitimate, or if not, defenseable, but that was the case 15 years ago in Türkiye, but isn't now. Things change.
(Note 1:"Dr. Bootlicker, the defendant wants the court to believe that she calmly placed herself between the agent and the minor he was trying to apprehend, and asserts that the agent's claim, that the defendant's actions constitute assault, is, in her words, 'ridiculous'. But am I correct in understanding that you view minutes 8 and 9 of the biometric data submitted to the court as characteristic of significant physical exertion that might be similar to that undergone by an assailant while commiting an assault?")
A couple of years ago Odeon turned our nearest theatre into a 'luxe' theatre (adult tickets £20), and the next nearest theatre was left as it was, but all tickets £5 each (tickets at both theatres where about £14 previously). I think it was an experiment to see which model was most economic: major investment in tech and comfort/£100+ for a family of four to watch a film with snacks and beverages/fewer tickets sold as a result OR minimal capex/far more affordable to attend/loads more tickets sold. The £5 tickets for all showings have stopped, but you can still get them a lot of the time (they have surge pricing for blockbuster releases,and some upgraded premium seating now). I think they've found a way to be affordable to the masses and fill seats, but still extract max revenue from better off families, by having half their theatres follow one model and half follow the other.
This is a bit off topic, but I occasionally used to sleep on the sofa in our first floor office in an old Georgian building in Fitzrovia. One occasion when I did that, I woke up at about 3.30 am with intense red light flooding through all the rear windows and the sound of loads of people chattering in the street out front, which seemed as busy as it normally would be in the daytime. I rushed to the front windows and looked down onto a street full of people, but all in 60s get up. I was still half asleep and panicked by the red light and it was totally disorientating to see a busy street of retro Londoners. I actually felt briefly nauseous but I went to the back windows and shaded my eyes, from the crazy glare from two arrays of red spotlights, which it turned out Edgar Wright was using to bounce light off our building, onto the cobbles below ...and began to understand what was going on. Was a relief to get a full explanation, for what had briefly felt like a weird time leap, when I went downstairs and chatted to the extras hanging around out front. The few seconds of woozy, confusion I spent in 1960s london seemed particularly appropriate when I saw Last Night in Soho a year or so later.
My wife used Apple Maps for a while here in the UK and driving in Europe. The results varied between amusing and traumatic. No issues ever with Google Maps since she swapped (but I know from experience it's not perfect). Apple maps would send her over tertiary roads through mountain passes that were snowed out, instead of salted/gritted primary roads, would show major highway junctions wildly (dangerously) inaccurately and showed areas that had lots of properly mettled roads as open countryside with no thoroughfares at all.
This is great, but also not well suited (in terms of visuals, name, language) to some of the audiences that need it most. A version that resonates more with middle aged men would be great. Oak or something.
About 10 years ago I quietly parked my Aunt's car in her garage so the driver's side door was about 6 inches away from the garage wall and got out of the passenger door. Although she insisted she was cognitively okay to drive, turned out she wasn't cognitively okay to work out how to get back into her car.
(Never mind, it's all in the training data now, will result in some competent, twee, but uncanny, slop).